On an earnings call full of talk about models and agents, the number that moved Snowflake’s stock was a count of customer accounts. The company said its AI coding assistant, Cortex Code, surpassed 9,100 accounts during its fiscal second quarter after adding more than 2,000 in the period, and that CoWork, an enterprise chatbot, expanded to 5,800 accounts. Investors took the adoption figures as evidence that artificial intelligence is finally showing up in Snowflake’s income statement, and they drove the shares up about 23 percent in after-hours trading Wednesday.
The report, released after the close, was strong by most measures. Total revenue for the quarter ended July 31 rose 35 percent from a year earlier to $1.55 billion, and product revenue climbed 37 percent to $1.49 billion, ahead of the roughly $1.48 billion that analysts had expected. Non-GAAP operating income rose about 70 percent to $237 million, lifting the operating margin to roughly 15 percent, and non-GAAP net income climbed about 99 percent to $235 million. On a GAAP basis the company still lost money, posting an operating loss of $263 million and a net loss of $192 million.
Management credited AI products for about half of the acceleration in growth. Customers adopting the assistants are also consuming more of Snowflake’s core data platform, executives said, which is the mechanism behind the accounts-and-revenue story: the tools pull more work into the warehouse they sit on. The company launched more than 330 product capabilities to general availability in the first half of the fiscal year, up 35 percent from a year earlier.
Snowflake also raised its forecast. The company lifted its full-year product revenue guidance to $6.07 billion from $5.84 billion, a target that implies growth of roughly 36 percent, and it nudged up its full-year non-GAAP operating margin view as well. Guidance that goes up alongside a quarter that beat estimates is the combination software investors are least used to seeing from this company.
The account numbers matter because enterprise software vendors have spent the AI boom promising future payoffs. Snowflake handed Wall Street something rarer: usage counts attached to named products, disclosed inside an earnings release, with a raised forecast attached. Analysts noted that the number of customers spending at least $1 million a year has more than doubled over the past three years, and that net revenue retention stood at 126 percent, meaning existing customers keep buying more.
The report is a turnabout for a company that spent two years explaining decelerating growth. Snowflake went public in September 2020 in the largest software listing of its time and became a pandemic favorite, only to see growth slow sharply as customers squeezed cloud budgets. Sridhar Ramaswamy, a former Google advertising chief who founded the search startup Neeva, took over as chief executive in 2024 and has pushed the company toward an AI-centered pitch built on the data warehouse that customers already run. CoCo and CoWork, both introduced this year, are the first products to carry that pitch directly, and the second quarter was the first in which the market could measure it.
There were softer spots beneath the headline numbers. Remaining performance obligations, a measure of contracted future revenue, came in at $9 billion, up 30 percent from a year earlier but below the roughly $9.5 billion that some analysts tracked. The AI products do not yet carry their own revenue line; the company points to platform consumption instead. And the stock had already risen about 40 percent this year before the report, so the after-hours jump to levels last seen in November 2021 puts a premium on the next few quarters of execution.
The margin story is part of the debate. Snowflake is still working toward GAAP profitability, and management has set a goal of reaching it by the fourth quarter of fiscal 2028. AI workloads carry real computing costs, which is why the company’s margin guidance is watched as closely as its revenue guidance, and the full-year non-GAAP product gross margin outlook of 74 percent was left unchanged.
The competitive picture sharpens the stakes. Databricks, the privately held rival that competes for the same data workloads, has been growing faster and commands a private-market valuation above most public software companies, and both firms are chasing the same agentic-AI business. Snowflake’s counterargument is governance: as AI agents gain access to corporate data, companies will want that work to happen inside a platform with controls they already trust, rather than across a patchwork of new tools.
For investors, the quarter answered one question and opened another. The answer is that AI adoption can accelerate Snowflake’s growth and its margins in the same quarter. The question is whether the account numbers keep climbing, whether usage converts into durable revenue, and whether the shortfall in performance obligations was timing or competition. Wednesday’s rally suggests the market is willing, for now, to believe the first half of that story.
